[Ecls-list] Swank ECL backend

[Geo Carncross]

On Jan 19, 2008 8:49 AM, Tobias C. Rittweiler <tcr at freebits.de> wrote:
> > Shouldn't make-weak-key-hash-table then use the default (non-weak)
> > implementation in swank-backend.lisp?
>
> This would result in a a quite severe memory leak, as all REPL results
> are stored in that hash-table.

That's true, but ECL first needs either weak hashtables or weak pointers to
really solve this.

In any event, Julian shouldn't be getting that error...